Project ideas from Hacker News discussions.

90% of US working hours are spent on tasks not covered well by AI benchmarks

📝 Discussion Summary (Click to expand)

Five prevalent themes in the discussion

  • Stakeholder management and office politics consume a large share of work
  • “About 80% of my software engineering job is fighting disagreements between stakeholders.” (qurren)
  • “50% office politics  30% stakeholder and expectations management  10% design/development.” (duttish)
  • “Stakeholders are a natural outcome of corporate evolution… inevitably results in suboptimal dynamics like bureaucracy, empire building…” (keeda)

  • Current AI benchmarks miss most real‑world tasks

  • “Our findings suggest that 90% of American working hours are spent on tasks with little or no coverage from existing benchmarks.” (10111001)
  • “I don’t know if it is 90%, but it is interesting how many things I do are not covered by standard tracking…” (qurren)
  • “The next little while is going to be very revealing for leadership around how little of the time involved in getting a solution to production is the coding.” (SketchySeaBeast)

  • Economic effects: wealth concentration, cost‑of‑living pressures, and distribution of gains

  • “Luxuries have gotten cheaper while basic necessities like food, housing, and medicine have become dramatically more expensive.” (stouset)
  • “They’d pay you nothing if they could.” (dclowd9901) – on AI‑driven labor cost reduction
  • “AI company leaders are so adamant about framing it as a post‑singularity utopia with UBI.” (dclowd9901)

  • Job automation impact is modest; productivity gains are slow‑growing

  • “A lot of what humans do does not have a tight feedback loop… productivity gains look a lot more like the very long‑term 2% per year.” (verdverm)
  • “We might be too ready to believe this is coming for everyone's job, not only generating code.” (ChrisLTD)
  • “So the problem is you can’t replace everyone halfway.” (Nevermark)

  • Physical/manual work and robotics remain hard for AI to replace in the near term

  • “But is it cheaper than paying someone around $5 in labor to do it? Because until it is, it's not economical to have a robot do it.” (devmor)
  • “Boston Dynamics has the best robot mechatronics in the world… they have been trying since 1992.” (robotresearcher)
  • “I mean, from my perspective as an engineer, electricians’ work is just extremely primitive.” (huijzer) – highlighting the gap between AI‑friendly tasks and skilled trades.

🚀 Project Ideas

StakeholderSignal

Summary

  • An AI assistant that ingests stakeholder comments from Slack, email, and meeting transcripts, detects disagreements, and proposes neutral, data‑driven resolutions.
  • Core value: cuts down time spent fighting stakeholder disagreements by providing an impartial “AI said so” recommendation that reduces ego‑driven pushback.

Details

Key Value
Target Audience Software engineers, product managers, and team leads who spend >30% of time aligning stakeholders
Core Feature Real‑time conflict detection + suggested compromise language backed by project goals and metrics
Tech Stack Python (FastAPI), LLM API (e.g., Llama 3), Vector DB (Weaviate), React frontend
Difficulty Medium
Monetization Revenue-ready: $12/user/month SaaS

Notes

  • HN commenters lament that 80% of engineering time is spent on stakeholder fights; this tool directly addresses that pain (qurren, duttish).
  • By providing an authoritative AI suggestion, it reduces the need for political posturing and endless back‑and‑forth.

FeedbackLoop AI

Summary

  • A desktop agent that passively watches IDE activity, git commits, and ticket updates to generate micro‑feedback loops for tasks that lack clear success signals.
  • Core value: creates the tight feedback needed for personal skill improvement and for generating high‑quality agentic training data on real work.

Details

Key Value
Target Audience Individual contributors (devs, writers, analysts) who feel their work lacks immediate feedback
Core Feature Automated micro‑reviews: diff‑level suggestions, productivity nudges, and skill‑gap highlights
Tech Stack Electron/Tauri, Language Server Protocol hooks, Whisper for voice notes, Postgres
Difficulty Medium
Monetization Revenue-ready: $8/month per pro user

Notes

  • verdverm notes many human activities lack tight feedback loops, hindering agentic workflows; this closes that gap.
  • HN users often wish for better self‑tracking (iugtmkbdfil834); the tool surfaces invisible work patterns.

BullshitJobDetector

Summary

  • Analyzes calendars, meeting transcripts, and task tags to estimate the proportion of time spent on low‑value, “bullshit” activities (e.g., status reporting, stakeholder appeasement).
  • Core value: makes invisible waste visible, enabling individuals and managers to target automation or process redesign.

Details

Key Value
Target Audience Engineers, managers, and HR leaders seeking to reduce unproductive work
Core Feature Time‑allocation dashboard with AI‑classified task categories (value‑add vs. overhead)
Tech Stack Node.js backend, Google Calendar API, Whisper speech‑to‑text, spaCy for classification, Grafana UI
Difficulty Low
Monetization Revenue-ready: $99/team/month

Notes

  • Multiple commenters reference “bullshit jobs” and the desire to eliminate meaningless stakeholder roles (m000, kenjackson).
  • Making the hidden overhead quantifiable aligns with HN’s love for data‑driven process improvement.

OfficePoliticsMitigator

Summary

  • A secure, anonymous survey and suggestion platform that surfaces political tensions, hidden agendas, and fief‑building behaviors, then offers mediated, game‑theoretic compromise paths.
  • Core value: reduces empire building, territorialism, and ego‑driven conflict by making politics visible and actionable.

Details

Key Value
Target Audience Mid‑size to large tech companies where office politics slows delivery
Core Feature Anonymous pulse surveys + AI‑generated mediation scripts based on stated incentives
Tech Stack Go microservices, PostgreSQL, React, LLM for mediation scripts
Difficulty Medium
Monetization Revenue-ready: $150/company/month (tiered by employee count)

Notes

  • HN threads repeatedly cite ego, fiefdom protection, and political battles as major time sinks (keeda, qurren, whstl).
  • By anonymizing feedback and providing structured compromise, it directly attacks the “too many cooks” problem.

HierarchyFlattener

Summary

  • An organizational‑design assistant that analyzes reporting lines, decision‑latency metrics, and performance‑review data to recommend flatter hierarchies and better‑aligned review cycles.
  • Core value: cuts unnecessary layers that cause misaligned incentives and slow feedback, addressing the “too many levels of hierarchy” pain point.

Details

Key Value
Target Audience CTOs, VPs of Engineering, and People Ops leaders in scaling tech firms
Core Feature Simulation‑driven org‑chart optimizer with impact estimates on decision speed and alignment
Tech Stack Scala (Akka) for graph algorithms, Python/Streamlit for UI, Neo4j for org graph
Difficulty High
Monetization Revenue-ready: $500/department/quarter (consulting‑style SaaS)

Notes

  • qurren identifies excessive hierarchy and performance‑review loops as a root cause of misalignment; this tool offers a concrete remedy.
  • HN users often dream of reducing layers to speed up decision making; this provides a data‑backed path.

Read Later